There was nothing found during the investigation to indicate that the aircraft suffered any mechanical malfunction prior to the crash. Based on the instrument analysis, the engine teardown results, the aircraft's speed during the approach and excursion over the ground, and the distance the aircraft travelled after the first impact, it was concluded that the engines were operating normally and were producing high power at the time of impact. The weather was good at the time of the approach, in that the ceiling was 4,000 feet above ground level (agl) and the visibility was 12 miles, and the crew flew a normal, uncomplicated ILS approach. In such conditions, the crew should have been able to successfully land the aircraft. The only apparent conditions that may have affected the final stages of the flight were the darkness and the drifting snow from the right tailwind. Radar and instrument indications show that the aircraft was set up for the ILS, and that the flight director system was engaged in the approach mode, although there is uncertainty as to whether the glide slope was engaged. The crew did not advise ATC of any problems with the aircraft; had they not been satisfied that they could make a safe landing, the crew would have commenced a missed approach procedure as the aircraft approached the runway. The pilot's adjustments of the ARCAL light system on final approach indicate that at least one of the pilots could see the runway lights, from about 3,000 feet agl and 6.6 miles from the threshold, and was adjusting their intensity. Analysis of the recorded radar data indicates that the aircraft, while on the approach, remained established on the localizer, correcting for the right quartering tailwind. The heading bug selection, 10 to the right of the inbound course, also indicates a correction for the wind. When the aircraft approached the runway threshold, its landing lights would have illuminated the drifting snow and the snow covering much of the runway surface, probably making it difficult to distinguish the runway's white centre line and, perhaps, the runway edge lights. The illuminated snow drifting across the runway at a 45 angle from behind the aircraft would give a pilot the illusion of lateral aircraft motion. Considering that there was no mechanical or aerodynamic explanation for a directional control problem, it is most likely that a flight control input, or lack of input, allowed the aircraft to drift to the left. This could have been the result of the pilot wanting to remain clear of the windrow on the right side of the runway, or his removal of the 10 crosswind correction in preparation for landing. The pilot's reference to the runway edge lights may have been degraded by the drifting snow, and when the aircraft began to drift to the left, in the same direction as the drifting snow, it could have been difficult for the pilot to detect and correct the aircraft's movement. Touching down in the snow off the left side of the runway would have surprised the pilot and would have affected his subsequent performance in the missed approach. Heading left off the runway, in the dark and with a lack of ground lights in that direction, the pilot had a limited horizon comprised of the snow surface illuminated by the aircraft's landing lights, which would have made recognition of the aircraft's attitude extremely difficult. The pilot's attention during the landing flare would have been concentrated on the visual environment outside of the aircraft, and it is likely that the pilot attempted to establish the missed approach attitude using outside references. That the aircraft was at various bank angles of wings level, 10 left, 45 left, and 10 left, until the ailerons jammed, indicates that the pilot had lost control of the aircraft during the missed approach attempt. Once the aileron jammed, the pilot could no longer control the bank of the aircraft. To maintain a nearly straight ground track of 261, the aircraft would have had to be banking back and forth along the entire track. The pilot also did not maintain the required nose-up pitch attitude; simulator trials and examination of the aircraft's climb performance demonstrated that the aircraft would have flown away from the ground had such an attitude been maintained. A pilot commencing a missed approach with reference to the cockpit instruments would normally select the go-around mode on the flight director so the V-bars could command the proper aircraft attitude on the attitude director indicator. Had the Gander ACC operational staff been aware of the RADEX capability to quickly identify an aircraft's last recorded radar position and had that information been provided to the agencies conducting the search, the aircraft crash site would have been located much sooner.Analysis There was nothing found during the investigation to indicate that the aircraft suffered any mechanical malfunction prior to the crash. Based on the instrument analysis, the engine teardown results, the aircraft's speed during the approach and excursion over the ground, and the distance the aircraft travelled after the first impact, it was concluded that the engines were operating normally and were producing high power at the time of impact. The weather was good at the time of the approach, in that the ceiling was 4,000 feet above ground level (agl) and the visibility was 12 miles, and the crew flew a normal, uncomplicated ILS approach. In such conditions, the crew should have been able to successfully land the aircraft. The only apparent conditions that may have affected the final stages of the flight were the darkness and the drifting snow from the right tailwind. Radar and instrument indications show that the aircraft was set up for the ILS, and that the flight director system was engaged in the approach mode, although there is uncertainty as to whether the glide slope was engaged. The crew did not advise ATC of any problems with the aircraft; had they not been satisfied that they could make a safe landing, the crew would have commenced a missed approach procedure as the aircraft approached the runway. The pilot's adjustments of the ARCAL light system on final approach indicate that at least one of the pilots could see the runway lights, from about 3,000 feet agl and 6.6 miles from the threshold, and was adjusting their intensity. Analysis of the recorded radar data indicates that the aircraft, while on the approach, remained established on the localizer, correcting for the right quartering tailwind. The heading bug selection, 10 to the right of the inbound course, also indicates a correction for the wind. When the aircraft approached the runway threshold, its landing lights would have illuminated the drifting snow and the snow covering much of the runway surface, probably making it difficult to distinguish the runway's white centre line and, perhaps, the runway edge lights. The illuminated snow drifting across the runway at a 45 angle from behind the aircraft would give a pilot the illusion of lateral aircraft motion. Considering that there was no mechanical or aerodynamic explanation for a directional control problem, it is most likely that a flight control input, or lack of input, allowed the aircraft to drift to the left. This could have been the result of the pilot wanting to remain clear of the windrow on the right side of the runway, or his removal of the 10 crosswind correction in preparation for landing. The pilot's reference to the runway edge lights may have been degraded by the drifting snow, and when the aircraft began to drift to the left, in the same direction as the drifting snow, it could have been difficult for the pilot to detect and correct the aircraft's movement. Touching down in the snow off the left side of the runway would have surprised the pilot and would have affected his subsequent performance in the missed approach. Heading left off the runway, in the dark and with a lack of ground lights in that direction, the pilot had a limited horizon comprised of the snow surface illuminated by the aircraft's landing lights, which would have made recognition of the aircraft's attitude extremely difficult. The pilot's attention during the landing flare would have been concentrated on the visual environment outside of the aircraft, and it is likely that the pilot attempted to establish the missed approach attitude using outside references. That the aircraft was at various bank angles of wings level, 10 left, 45 left, and 10 left, until the ailerons jammed, indicates that the pilot had lost control of the aircraft during the missed approach attempt. Once the aileron jammed, the pilot could no longer control the bank of the aircraft. To maintain a nearly straight ground track of 261, the aircraft would have had to be banking back and forth along the entire track. The pilot also did not maintain the required nose-up pitch attitude; simulator trials and examination of the aircraft's climb performance demonstrated that the aircraft would have flown away from the ground had such an attitude been maintained. A pilot commencing a missed approach with reference to the cockpit instruments would normally select the go-around mode on the flight director so the V-bars could command the proper aircraft attitude on the attitude director indicator. Had the Gander ACC operational staff been aware of the RADEX capability to quickly identify an aircraft's last recorded radar position and had that information been provided to the agencies conducting the search, the aircraft crash site would have been located much sooner. The flight crew was certified and qualified for the flight in accordance with existing regulations. The St. John's FSS operator did not have the actual Stephenville wind direction and speed. The wind velocity he passed to the crew was from the latest Stephenville observation and was 040 at 17 knots, within the tailwind landing limitations of the aircraft. The actual Stephenville wind of 040 magnetic at 20 knots with gusts to 22 knots exceeded the aircraft's maximum allowable tailwind component for landing. The pilot attempted a missed approach after the aircraft had touched down in the snow, just off the runway surface. The pilot did not maintain the correct aircraft attitude for a missed approach. The pilot did not select the go-around mode on the flight director during the missed approach. The CVR cockpit area microphone channel was not recorded for undetermined reasons. The capabilities of RADEX to quickly locate a missing aircraft were not known to the Gander ACC operational management staff. All major aircraft components were identified at the wreckage site, and no mechanical malfunction was identified. Engine and instrument analysis identified that both engines were operating at high power during the impact. There was an ELT installed in the aircraft, although, according to FARs, the aircraft was not required to be so equipped. The ELT did not activate at impact; the batteries were not the correct type and were overdue for replacement.Findings The flight crew was certified and qualified for the flight in accordance with existing regulations. The St. John's FSS operator did not have the actual Stephenville wind direction and speed. The wind velocity he passed to the crew was from the latest Stephenville observation and was 040 at 17 knots, within the tailwind landing limitations of the aircraft. The actual Stephenville wind of 040 magnetic at 20 knots with gusts to 22 knots exceeded the aircraft's maximum allowable tailwind component for landing. The pilot attempted a missed approach after the aircraft had touched down in the snow, just off the runway surface. The pilot did not maintain the correct aircraft attitude for a missed approach. The pilot did not select the go-around mode on the flight director during the missed approach. The CVR cockpit area microphone channel was not recorded for undetermined reasons. The capabilities of RADEX to quickly locate a missing aircraft were not known to the Gander ACC operational management staff. All major aircraft components were identified at the wreckage site, and no mechanical malfunction was identified. Engine and instrument analysis identified that both engines were operating at high power during the impact. There was an ELT installed in the aircraft, although, according to FARs, the aircraft was not required to be so equipped. The ELT did not activate at impact; the batteries were not the correct type and were overdue for replacement. Shortly after crossing the runway threshold, the aircraft began moving to the left of the runway. The motion probably was undetected by the pilot until the aircraft touched down off the left side of the runway surface. The pilot did not maintain the proper aircraft attitude during an attempted missed approach, and the aircraft struck the terrain.Causes and Contributing Factors Shortly after crossing the runway threshold, the aircraft began moving to the left of the runway. The motion probably was undetected by the pilot until the aircraft touched down off the left side of the runway surface. The pilot did not maintain the proper aircraft attitude during an attempted missed approach, and the aircraft struck the terrain. The capability of RADEX to quickly locate the last radar position of missing or overdue aircraft was recognized by NAV CANADA authorities and action was taken to make the equipment and the program available to operations personnel in all ACCs for use in similar occurrences. NAV CANADA conducted training sessions for Data Systems Controllers in all ACCs specifically aimed at the use of RADEX as a search tool. As a result of this accident and in an effort to enhance the safety of operations at Stephenville airport, the Airport Authority initiated discussions with NAV CANADA and Transport Canada to establish an Authorized Approach Unicom (AAU) service to provide operational information to pilots for the purpose of conducting instrument approaches published in the Canada Air Pilot (CAP). An AAU is authorized to provide airport advisory services including surface wind speed and direction, current altimeter setting and runway condition (surface condition, vehicles, etc.) to aircraft.Safety Action Taken The capability of RADEX to quickly locate the last radar position of missing or overdue aircraft was recognized by NAV CANADA authorities and action was taken to make the equipment and the program available to operations personnel in all ACCs for use in similar occurrences. NAV CANADA conducted training sessions for Data Systems Controllers in all ACCs specifically aimed at the use of RADEX as a search tool. As a result of this accident and in an effort to enhance the safety of operations at Stephenville airport, the Airport Authority initiated discussions with NAV CANADA and Transport Canada to establish an Authorized Approach Unicom (AAU) service to provide operational information to pilots for the purpose of conducting instrument approaches published in the Canada Air Pilot (CAP). An AAU is authorized to provide airport advisory services including surface wind speed and direction, current altimeter setting and runway condition (surface condition, vehicles, etc.) to aircraft.